Re: Disable HTT on Serverworks GC-HE chipset

Hi Scott,

I have tried without HTT because someone on #bsdcode told me
to do so (ket_). Maybe you know who he is :-) Not all revisions
of the chipset seem to be buggy so far, but rev. 0 is for sure.

http://www.plasma-online.de/bilder/identify/chips/serverworks/ss3_ws_nb6536_p01.jpg
http://www.plasma-online.de/bilder/identify/chips/serverworks/nb6536_p02.gif

Serversboards with this chipset are:

- Dell PowerEdge 2650
- IBM xSeries
- Supermicro E7500

CNB20-HE (NB6536 - ServerSet III HE Northbridge 2.0)
[0008] CPU - PCI bridge
[0013] CPU - PCI bridge
[0009] CPU - PCI/AGP bridge
[0009.cc_0600] CPU - PCI bridge (default)
[0009.cc_0604] CPU - AGP bridge (when used with NB6555AGP CIOB)

> <ket_> mbr - i've had a LOT of people tell me the GC-HE w/HTT causes unknown
>  lockups, unreproducible errors, memory errors, and strange bus errors. on
>  multiple operating systems.

> <ket_> GC-HE itself is acknowledged to be a buggy chipset, they're up to
  something like rev8 on the silicon.

>> Have you determined that the instability is solely due to HTT
>> in particular and not SMP in general?  In other words, do you
>> have multiple CPUs in the system, or is it just a single CPU
>> with HTT?  Also, do you have the latest BIOS for your motherboard?
>> What stepping are your CPUs?

The box has 2 CPU's. Without HTT it is rocking stable. And yes, the
X-Series 345 have the latest Bios available. This dmesg is without
HTT.

CPU: Intel(R) Xeon(TM) CPU 3.06GHz (3059.98-MHz 686-class CPU)
  Origin = "GenuineIntel"  Id = 0xf27  Stepping = 7

Features=0xbfebfbff<FPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CLFLUSH,DTS,ACPI,MMX,FXSR,SSE,
SSE2,SS,HTT,TM,PBE>
real memory  = 2147397632 (2047 MB)
avail memory = 2100260864 (2002 MB)
ACPI APIC Table: <IBM    SERONYXP>
FreeBSD/SMP: Multiprocessor System Detected: 2 CPUs
 cpu0 (BSP): APIC ID:  0
 cpu1 (AP): APIC ID:  6
ioapic2 <Version 1.1> irqs 32-47 on motherboard
ioapic1 <Version 1.1> irqs 16-31 on motherboard
ioapic0 <Version 1.1> irqs 0-15 on motherboard
Pentium Pro MTRR support enabled

On a other box with HTT still enable it looks like:

CPU: Intel(R) Xeon(TM) CPU 3.06GHz (3059.98-MHz 686-class CPU)
  Origin = "GenuineIntel"  Id = 0xf29  Stepping = 9

Features=0xbfebfbff<FPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CLFLUSH,DTS,ACPI,MMX,FXSR,SSE,
SSE2,SS,HTT,TM,PBE>
  Hyperthreading: 2 logical CPUs
real memory  = 2147397632 (2047 MB)
avail memory = 2100260864 (2002 MB)
ACPI APIC Table: <IBM    SERONYXP>
FreeBSD/SMP: Multiprocessor System Detected: 4 CPUs
 cpu0 (BSP): APIC ID:  0
 cpu1 (AP): APIC ID:  1
 cpu2 (AP): APIC ID:  6
 cpu3 (AP): APIC ID:  7
ioapic2 <Version 1.1> irqs 32-47 on motherboard
ioapic1 <Version 1.1> irqs 16-31 on motherboard
ioapic0 <Version 1.1> irqs 0-15 on motherboard
Pentium Pro MTRR support enabled


You can see the chipset here:

pciconf -lv

hostb0_at_pci0:0:0:        class=0x060000 card=0x00000000 chip=0x00141166 rev=0x33
hdr=0x00
    vendor   = 'ServerWorks (Was: Reliance Computer Corp)'
    device   = 'CNB20-HE Host Bridge'
    class    = bridge
    subclass = HOST-PCI
hostb1_at_pci0:0:1:        class=0x060000 card=0x00000000 chip=0x00141166 rev=0x00
hdr=0x00
    vendor   = 'ServerWorks (Was: Reliance Computer Corp)'
    device   = 'CNB20-HE Host Bridge'
    class    = bridge
    subclass = HOST-PCI
hostb2_at_pci0:0:2:        class=0x060000 card=0x00000000 chip=0x00141166 rev=0x00
hdr=0x00
    vendor   = 'ServerWorks (Was: Reliance Computer Corp)'
    device   = 'CNB20-HE Host Bridge'
    class    = bridge
    subclass = HOST-PCI


While using google I found the GC-LE chipsets seems to have problems
with write-coalescing:

>      ServerWorks LE chipsets have problems with write-combining
+      Don't allow it and leave room for other chipsets to be tagged

Have we a workaround for this too in our source ?

Martin